Since we are arriving at the end of the year, I have taken the mission to replicate data between SQL server and TERADATA. The worst part in this task, is to install and configure a Goldengante in a Windows environment.
Believe, it is not possible to do a Unix installation of goldengate to collect data from SQLserver, goldengate binary needs to be installed on Windows SQLserver host.
After installing the GG binaries, it is good practice to add the MGR as a Windows service:
C:\goldengate> install addevents addservice manualstart Oracle GoldenGate messages installed successfully. Service 'GGSMGR' created. Install program terminated normally.
In order for GG to access the sql database, you need to create a data source (ODBC), and configure a new system DSN (here is db0sql1), and select SQL Server as the database driver.
To perform a DBLOGIN:
DBLOGIN SOURCEDB db0sql1, USERID ggate, PASSWORD ??????
To configure extract uses the same mode of other GG process:
ADD EXTRACT E_SQL, TRANLOG, BEGIN NOW ADD EXTRAIL ./dirdat/tr, EXTRACT E_SQL, MEGABYTES 100 ADD TRANDATA dbo.DLOG_ERRORS allcols
GGSCI (sqlserverdb) 2>edit param e_sql EXTRACT E_SQL TRANLOGOPTIONS MANAGESECONDARYTRUNCATIONPOINT SOURCEDB db0sql1, USERID ggate, PASSWORD ?????? CACHEMGR CACHESIZE 1GB exttrail ./dirdat/tr --TABLE MAP TABLE dbo.DLOG_ERRORS; TABLE dbo.SAC_DATA; TABLE dbo.SAC_LIST; TABLE dbo.SAC_TITLE;
It is possible to configure an extract process sending trails directly to the destination, which in this case would be the GG-TERADATA, but we always configure a pump process as good practice, because in the event of any communication problem, it will not affect the extraction process.
The Pump and Replicat process configuration, will be presented in Part 2.